NEW YORK (GenomeWeb News) – A three-gene signature can distinguish indolent from aggressive prostate cancers, researchers led by Cory Abate-Shen, a professor of urological oncology at Columbia University Medical Center, reported in Science Translational Medicine today.
Of the approximately 238,600 prostate cancers that are diagnosed each year in the US, most are low-risk and slow growing. Currently, that risk is determined through Gleason grading, which relies on a histopathological assessment of a biopsy.
To avoid over-treatment — prostate cancer is often found in its early stages — clinicians typically employ watchful waiting or active surveillance to monitor men with low-risk tumors. However, Abate-Shen and her colleagues noted that such an approach may cause doctors and their patients to miss an opportunity to treat prostate cancers that appear to be low risk, but that actually are aggressive. To differentiate between those two classes, they developed a three-gene signature that they said could become a part of the monitoring process.
"The problem with existing tests is that we cannot identify the small percentage of slow-growing tumors that will eventually become aggressive and spread beyond the prostate," author Mitchell Benson, a urology professor at Columbia, said in a statement.
To develop their signature, the researchers posited that, as prostate cancer is a disease of aging, aggressive and indolent forms of the disease might differ in their expression of genes involved in aging and senescence.
Abate-Shen, Benson, and their colleagues first developed a 377-gene signature, mined mainly from a meta-analysis of aging genes. It was, they noted, enriched for pathways associated with aging-related diseases, though not with protumorigenic pathways.
Using a gene-set enrichment analysis, they determined that that set of genes was enriched for genes that were down-regulated in aggressive prostate cancer by evaluating those genes' expression in a dataset containing aggressive prostate cancer and matched, adjacent controls. Additionally, a meta-analysis of down-regulated genes in aggressive prostate, lung, and breast cancers identified a set of 68 down-regulated genes that were enriched in aggressive tumors.
To determine which subset of the original 377-gene signature was enriched in indolent prostate cancer, the researchers turned to a mouse model system, as a dataset of purely indolent cancers in humans was not available. The prostate lesions in indolent mice had up-regulated genes in the original gene signature, as compared to control mice.
By combining those findings, Abate-Shen and her colleagues developed a set of 19 genes that are up-regulated in indolent prostate cancers and down-regulated in aggressive prostate cancers, a set of genes they dubbed an indolence signature.
They validated the 377-gene signature — which included the 19 genes of the indolence signature — in a dataset of prostate cancer samples comprising a range of Gleason scores, though they particularly focused on low Gleason scores as well as time to biochemical recurrence. From their gene-set enrichment analyses, the researchers determined that the differential enrichment of genes in the aging and senescence signature could distinguish low-scoring tumors that will remain indolent from those that will become aggressive.
Using a decision tree learning algorithm based on the 19-gene indolence signature, the researchers homed in on a smaller set of genes — FGFR1, PMP22, and CDKN1A — that appeared to act as prognostic biomarkers and that would be even better at distinguishing aggressive and indolent prostate cancers.
The presence of FGFR1 in an indolence signature was "rather unexpected," the researchers said, as it encodes a receptor for fibroblast growth factor signaling in the prostate and plays a role in prostate development and prostate tumorigenesis
The inclusion of CDKN1A, by contrast, was not surprising as it is a cell cycle regulatory gene whose down-regulation has been associated in the past with cancer progression.
PMP22, surprisingly, encodes a glycoprotein that makes up a portion of the myelin protein in the nervous system, and its inclusion in the gene signature suggests that it might have a "new and unexpected role" in the prostate, the researchers noted.
At both the mRNA and protein levels, the researchers found that this set of three genes could stratify low Gleason score prostate tumors into high- and low-risk groups and could do it better than the 19-gene indolence signature.
To test the three-gene signature in a clinical environment, the researchers conducted a blinded retrospective analysis of patients who presented at Columbia's medical center between 1992 and 2012 with clinically low-risk prostate cancer. The patients were monitored, receiving repeat PSA testing as well as repeat biopsies.
The researchers compared expression levels of FGFR1, PMP22, and CDKN1A in the initial biopsies from 14 patients who failed surveillance with those of 29 patients who did not fail surveillance over a 10-year period. Failing surveillance entailed seeing an increase of cancer grade or volume on biopsy.
There was, the researchers wrote, "marked correlation between the expression of FGFR1, PMP22, and CDKN1A and outcome."
"The bottom line is that, at least in our preliminary trial, we were able to accurately predict which patients with low-risk prostate cancer would develop advanced prostate cancer and which ones would not," Abate-Shen said.
The researchers suggested that their signature be further evaluated to determine whether it should incorporated into active surveillance protocols. "When combined with other clinical parameters, such a prognostic test performed on biopsies may contribute to more effective monitoring of men on active surveillance protocols," they added.
A number of the study's authors, including Abate-Shen and Benson, have a patent pending for the use of the signature for cancer screening.